fix: keep overlap category names from element 100 on

CHANGELOG.md

@@ -1,7 +1,9 @@
-## 0.27.27-dev0
+## 0.27.27-dev1
 
 ### Fixes
 
+- **Keep overlap category names intact from element 100 on.** `identify_overlapping_or_nesting_case()` split `{ix}. {category}` labels with a 3-character slice, so from index 100 the leftover digits leaked into `overlapping_elements` and `overlapping_case`. It now splits on the first `". "`.
+
 - **Preserve serialized element identity when rehydrating JSON.** `partition_json()` no longer
   replaces element IDs or stored metadata with attributes from the JSON container. Explicit
   `metadata_filename`, `metadata_last_modified`, `url`, and `text_as_html` overrides remain supported,

test_unstructured/test_utils.py

@@ -7,7 +7,7 @@
 
 from unstructured import utils
 from unstructured.documents.coordinates import PixelSpace
-from unstructured.documents.elements import ElementMetadata, NarrativeText, Title
+from unstructured.documents.elements import ElementMetadata, NarrativeText, Table, Title
 
 
 @pytest.fixture()
@@ -312,6 +312,43 @@ def test_catch_overlapping_and_nested_bboxes_non_overlapping_case():
     assert overlapping_cases == []
 
 
+@pytest.mark.parametrize("n_filler", [0, 100, 1000])
+def test_catch_overlapping_and_nested_bboxes_keeps_category_past_element_99(n_filler):
+    coordinate_system = PixelSpace(width=1000, height=1000)
+    fillers = [
+        NarrativeText(
+            text=f"filler {i}",
+            coordinates=((0, i), (0, i + 1), (5, i + 1), (5, i)),
+            coordinate_system=coordinate_system,
+            metadata=ElementMetadata(page_number=1),
+        )
+        for i in range(n_filler)
+    ]
+    elements = fillers + [
+        Table(
+            text="cell a cell b",
+            coordinates=((0, 0), (0, 500), (500, 500), (500, 0)),
+            coordinate_system=coordinate_system,
+            metadata=ElementMetadata(page_number=2),
+        ),
+        Title(
+            text="cell a",
+            coordinates=((10, 10), (10, 50), (100, 50), (100, 10)),
+            coordinate_system=coordinate_system,
+            metadata=ElementMetadata(page_number=2),
+        ),
+    ]
+
+    overlapping_flag, overlapping_cases = utils.catch_overlapping_and_nested_bboxes(elements)
+
+    assert overlapping_flag is True
+    assert overlapping_cases[0]["overlapping_elements"] == [
+        f"Table(ix={n_filler})",
+        f"Title(ix={n_filler + 1})",
+    ]
+    assert overlapping_cases[0]["overlapping_case"] == "nested Title in Table"
+
+
 def test_only_returns_singleton_iterable():
     singleton_iterable = [42]
     result = utils.only(singleton_iterable)

unstructured/__version__.py

@@ -1 +1 @@
-__version__ = "0.27.27-dev0"  # pragma: no cover
+__version__ = "0.27.27-dev1"  # pragma: no cover

unstructured/utils.py

@@ -500,8 +500,9 @@ def identify_overlapping_or_nesting_case(
     type1, type2 = label_pair
     ix_element1 = "".join([ch for ch in type1 if ch.isnumeric()])
     ix_element2 = "".join([ch for ch in type2 if ch.isnumeric()])
-    type1 = type1[3:].strip()
-    type2 = type2[3:].strip()
+    # Labels are "{ix}. {category}", and ix can have any number of digits.
+    type1 = type1.split(". ", 1)[-1]
+    type2 = type2.split(". ", 1)[-1]
     box1_corners = _convert_coordinates_to_box(box1)
     box2_corners = _convert_coordinates_to_box(box2)
     x_bottom_left_1, y_bottom_left_1, x_top_right_1, y_top_right_1 = box1_corners